For a small project, I have to calculate a few things via SPSS. However, I am not sure which tests to use.
I have multiple variables. Independent variable: categorical variable with two levels. A group with music and a group without music (music versus no music). Dependent variable: continuous variable. This is a score on a test. Mediator variable: three questions (7-points Likert scale) which together represent the mediator disfluency.
I want to check if music has an effect on the score. Disfluency is a mediator because (shortly) this is the experience of ease or difficulty regarding finishing a mental task. If the disfluency is higher, the score is probably higher. The background music causes disfluency.
I wanted to do an independent t-test but I cannot add the mediator. Can I use the results of an independent t-test for a mediator analysis? Can someone help me with this?
Thank you in advance.
If you want to check if the DV has an effect on the IV, why do you want to include a mediator (or three)? Subsequent research questions, like is the effect of the IV on the DV mediated by these would require this. Second, the choice will depend on what the test is like. Is it a test compose of multiple items where they can get the answer right or wrong? Depending on sample size (you said small project ... how small?) item response theory (you may have run across this as latent trait models) could be helpful. It is usually wise not to jump in and immediately calculate the percentage correct (though this may end up being a good enough measure).
Lisa Neve. I'm not sure whether I have a good grasp on the nature of your data or research question (Daniel W. might have a much better idea), but I wonder whether a 2-way analysis of variance might be the way to go. You could look for main and interaction effects.
Or maybe consider multiple regression?
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS